Gene function prediction through logic analysis of gene associations

A major challenge facing biological research is the assignment of a function to uncharacterized genomic sequences. I will describe an approach for the functional characterization of gene products, which combines high-throughput bioinformatics and low-throughput, targeted experiments. The prediction of gene functions is based on the logic analysis of the evolutionary relationships existing between known and unknown genes involved in the same metabolic process. Bioinformatic predictions are validated in tageted experiments through the structural and functional characterization of gene products. Examples will be presented illustrating the application of such method to the study of the enzymes involved in purine catabolism. However, the method has a general applicability in the basic task of finding a function for unknown genes